Optimal organization of musical
playlists for physical exercising
Karolien Dons
Pro Gradu Thesis
Master’s programme in Music, Mind and Technology
Faculty of Humanities
University of Jyväskylä
June 2009
JYVÄSKYLÄN YLIOPISTO
Tiedekunta – Faculty
Laitos – Department
Faculty of Humanities
Music Department
Tekijä – Author
Karolien Sofie Katrien Dons
Työn nimi – Title
Optimal Organization of Musical Playlists for Physical Exercising
Oppiaine – Subject
Työn laji – Level
Music, Mind and Technology
Master
Aika – Month and year
Sivumäärä – Number of pages
June 2009
59
Tiivistelmä – Abstract
Research has proven that listening to music during physical activity can result in a better
performance quality. This thesis proposes the concept of Optimal Organization. It
suggests that next to a correct song choice, the selection of tracks can be organized
optimally, so that also the order and timing of the songs affect motivation and
performance quality of physical exercising. The concept consists of a collection or
grammar of preconditions helping the music ordering towards optimal organization. The
preconditions can be divided into three categories: internal musical factors, extra-musical
factors and non-musical factors.
The theory of optimal organization is tested empirically in a highly ecological and
exploratory experiment whereby participants run alternately with optimally and nonoptimally asynchronously organized music as guidance. Physiological and psychological
variables are recorded respectively during and after the performance executions. Data
analysis shows that heart rate and general performance feeling increased, and that energy
expenditure, length of the exercise (in time) and perceived exertion decreased in the cases
where the optimal playlist was listened to.
From the studies’ results a prediction model is extracted plotting the mechanism of
optimal playlist use in physical activity.
Asiasanat – Keywords
music & sports, ergogenics, flow, musical playlists
Säilytyspaikka – Depository
Muita tietoja – Additional information
1
Table of contents
List of tables................................................................................................................................3
List of Illustrations......................................................................................................................4
1. Introduction.....................................................................................................................5
2. Optimal Organization: background and theory...............................................................8
2.1 Music’s ergogenic quality.........................................................................................8
2.2 The Brunel Music Rating Inventory........................................................................10
2.3 Towards a music selection system for physical exercising.....................................12
2.4 Optimal Organization.............................................................................................14
3. Experiment....................................................................................................................19
3.1 Research questions and hypothesis.........................................................................19
3.2 Experiment’s design…............................................................................................21
3.2.1
Participants.................................................................................................21
3.2.2
Procedure....................................................................................................21
3.2.2.1 Prior to trials..........................................................................................21
3.2.2.2 Building the optimal playlist: two examples.........................................23
3.2.2.3 Execution of trials..................................................................................24
3.2.3
Instruments..................................................................................................26
4. Analysis & Results........................................................................................................28
4.1 Data analysis...........................................................................................................28
4.2 Physiological responses: heart rate data................................................................29
4.3 Spent calories..........................................................................................................32
4.4 Length of the exercise.............................................................................................33
4.5 Affective responses..................................................................................................34
4.5.1
RPE.............................................................................................................34
4.5.2
General feeling............................................................................................35
4.5.3
RPE vs. general feeling...............................................................................36
2
4.5.4
Mean HR vs. RPE.......................................................................................37
4.6 Modelling optimal playlist behaviour.....................................................................40
5. Conclusion....................................................................................................................43
5.1 Limitations of the experimental design...................................................................43
5.2 Towards a music selection system for physical exercising.....................................45
5.3 Enhancement of the performance quality?.............................................................46
References.................................................................................................................................47
Links.........................................................................................................................................50
Attachments..............................................................................................................................51
Attachment A: Overview preconditions of Optimal Organization...............................51
Attachment B: Preceding questionnaire for the participants........................................52
Attachment C: Evaluation form of the exercise...........................................................56
3
List of tables
Table 1: Optimal playlist participant 1
Table 2: Optimal playlist participant 6
Table 3: Repeated measures ANOVA results for all participants testing differences between
the four performance trials
Table 4: Related t-test results for all participants testing the differences between optimal and
random playlist conditions
Table 5: Energy expenditure for optimal and random trials (in Kcal)
Table 6: Exercise durations (in seconds)
Table 7: RPE data for the six participants’ optimal and random trials including the mean of
optimal and random
Table 8: General feeling data for the six participants’ optimal and random trials
Table 9: Participants’ dependent variable overview
4
List of illustrations
Figure 1: Training scheme
Figure 2: Mean heart rate and RPE for the exercises Optimal 1 & 2 and Random 1 & 2
Figure 3: Predictive model for optimal playlist use in sports setting
5
1. Introduction
At the London Marathon of 2008, besides the running the Run To The Beat experiment was
set up: 17 dj’s and bands playing music on the circuit for the participants. The songs’ beats
per minute were calculated scientifically beforehand as being highly motivating at that
specific moment of the marathon’s track.
This is just one of several examples illustrating the use of music as accompaniment for
individual sport exercising. Surfing on the trend of combining health and pleasure in leisure
consumption, using music is a growing phenomenon both in team sports and individually
performed athletic activities.
The arrival of compact portable music players has made it easy to reach for music during
training or competition. But also marketing has found the potential of the tandem music and
sports. Examples of this are the numerous training devices incorporating music player
functions such as Adidas’ MiCoach, Nike+iPod or Yamaha’s BODiBEAT. Also big sports
events programming music such as the aforementioned Run To The Beat project sidekicking
the London Marathon illustrate this new trend in the management of free time activities. A
last example of a successful event is Sporza Musica organized by the Flemish classical radio
Klara). Under the slogan “Classical music makes you move!” sports activities guided by
(partially live played) classical music were programmed.
But there is more than just finding fun in listening to music while exercising. Research has
proven that listening to music during physical activity can result in a better performance
quality; in the sense that the performance benefits from the music. Music elicits long-term and
short-term effects on physical training (Karageorghis et al., 1999). Long-term effects are
heightened work-rate and endurance; three short-term psychophysical effects produced are
improvement of mood, reduction of perception of exertion and arousal control (Karageorghis
et al., 1999, 714-5). The first two can be facilitated through music listening before and/or
during the exercise; exertion reduction appears as an effect of listening during the physical
effort. Some attempts to sketching and investigating the possibilities that music offers for a
workout have been made, but nevertheless a lot of investigation needs to be done still since
6
the field lacks theoretical bases and suffers from methodological limitations (Karageorghis &
Terry, 1997).
The use of music in sports is an example of what DeNora calls a prosthetic technology (2000,
102); it extends the body by affording capacity, motivation, co-ordination, energy and
endurance. DeNora assigns music as having the power to device bodily coordination and
organization.
There are five ways in which music aids athletic performance (Karageorghis, 1999;
Karageorghis & Priest, 2008).
- Firstly music can distract the mind from sensations of fatigue and therefore lower the
perception of exertion. This phenomenon is called dissociation (1) and works at best at
submaximal workout intensity.
The use of personal music players can intensify this
experience, since the in-ear device provides direct stimuli and blocks out unwanted thoughts
of physical pain or exhaustion (Bull, 2000).
- Secondly music helps regulating arousal (2) and can therefore be used to bring the athlete in
the right mood prior to training or performance (North & Hargreaves, 2000; Bishop et al.,
2007), but evenly during physical activity which is the focus in this study. Depending on
personal need and/or preference psyching up or calming down is being achieved. This
strategy runs equally with the usage of imagery in the enhancement of sport performance
(Weibull, 2008).
- In exercising, a repetitive physical movement synchronization (3) with the music can be
aiding the performance efficiency too. Music provides temporal cues and can exactly be
matched with the target tempo of the exerciser.
- Music affects the acquisition of motor skills positively (4) (Karageorghis & Priest, 2008). In
contrast with environmental sounds of street life - no matter how loud or chaotic -, songs are
able to produce the configuration or force to permit the user to move correctly (Bull, 2000).
This phenomenon is employed very often in the refinement of physical movements.
- Lastly, a consequence of using the correct music can be the attainment of flow (5), a state of
complete optimal functioning of body and mind on auto-pilot with minimal conscious effort
(Csíkszentmihályi, 1990).
State of flow is for many athletes the ultimate provision for
optimal performance and is commonly associated with a positive psychological state and
enjoyment of immersion in the activity. Choosing the right music is a key factor in reaching
or maintaining flow (Karageorghis, 1999). Again the use of portable players can enhance this
experience through facilitating the feeling of centeredness. Bull (2000) describes it as being
7
‘cocooned’; structuring and separating the outside world and fully experience the inside
world. Craig (in: Acevedo et al. 2006) defines a similar concept under the term homeostasis
and sees herein maintaining balance between affective components and physiological
condition of the body as the main goal. Changes in a particular condition “usually affect
several aspects, and they elicit a combination of homeostatic responses that serve to restore
an optimal balance of conditions” (Craig in: Acevedo et al., 2006, 15).
Let this quotation be the starting point of the coming up research report. Affective
components and physiological components do demand an optimal balance in exercising.
Hence its context of exercising is an unstable environment, working load, exertion and
feelings vary during the scope of an exercise. Maintaining the optimal balance is therefore a
challenge. This research investigates music as a mediator to facilitate this balance.
The present research commences with an overview of the existing knowledge on the topic of
music in sports in part 2. It sketches the general field and trims down the sphere focus
towards playlists for exercising. Ending up at the proposition of a theoretical design of
optimal playlists organization, the theoretical background is round off.
Part 3 consists of the empirical part of this research and explains the experiment carried out to
test optimal organization. Starting with hypotheses and matching research questions, the
design of the experiment (including participants, chronological description of the procedure
and material) and steps in data analysis are included in this chapter.
The actual analysis of the data and significant outcomes and results are interpreted in part 4.
The results cumulate in the proposal of a predictive model of optimal organization of musical
playlists.
At last part 5 summarizes the outcomes and overviews the research considering the
limitations of the investigation and possible future directions.
After the conclusion references and attachments are listed.
8
2.
Optimal Organization:
background and theory
2.1 Music’s ergogenic quality
That music can extend endurance (Crust, 2004; Crust & Clough, 2006) and enhance
motivation (Karageorghis, Jones & Low, 2006; Karageorghis et al. 2009; Simpson &
Karageorghis, 2006) more than a situation without music (or white noise condition, Crust
2004), has been shown consistently.
A large part of the music and sports research has sought the musical factors 1 responsible for
facilitating the ergogenic effect on physical activity, meaning that music enhances the
exercise quality.
It is generally accepted that the most influential aspects are tempo, rhythm and rhythmic
pattern(s) (Iwanaga, 1995; Szabo, Small & Leigh, 1999; Crust & Clough, 2006; Edworthy &
Waring, 2006; Karageorghis, Jones & Low, 2006; Simpson & Karageorghis, 2006; Priest &
Karageorghis, 2008). Empirical studies have not defined yet whether musical tempo and/or
beat have to be in synchronization with the exercise movements for a higher performance
quality. Synchronous music is suggested to have the most positive effect on non-elite
sportsmen (according to Simpson & Karageorghis, 2006 and Karageorghis et al., 2009)
whereas no evidence for the deteriorating effect of listening to asynchronous music – music
played in the background without any conscious effect of the performer to stay in time – has
been found (Crust & Clough, 2006). Karageorghis and Terry (1997) however pointed out that
humans possess a predisposition to respond to musical tempo via bodily movement. Iwanaga
(1995) has suggested that tempo preference for everyday activity is situated within the range
of normal heart rate range.
1
For an overview of music’s ergogenic qualities see Attachment A.
9
Furthermore it is assumed that harmonic and melodic aspects affect the quality of the
performance (Karageorghis et al., 1999, 2006; Karageorghis & Priest, 2008). It is said for
instance that a highly pitched melody with clarity in harmony brings the sportsperson to
exercise more intensively. Especially female sung melodies facilitate this. Melody and
harmony are not widely studied with regard to the topic of music and sports; nor are there
enough clear results to speak about factual knowledge.
Apart from qualities from the music itself, also extra-musical and non-musical factors affect
the performance. Familiarity (Crust, 2004b; Karageorghis et al., 1999; Karageorghis & Priest,
2008) and musical preference (Crust & Clough, 2006; Karageorghis et al., 1999;
Karageorghis & Priest, 2008) are said to be decisive in a sportsman’s playlist choice. Cultural
and personal associations to sport, movement or triumph can be significant likewise
(Karageorghis et al., 1999, 2006; Karageorghis & Priest, 2008), as also the type or genre of
music (Tenenbaum et al., 2004) can be influential although very often both in and outside
scientific setting is chosen for music that appeared in popular charts.
10
2.2 The Brunel Music Rating Inventory
None of the above mentioned researches could allocate the exclusive decisive musical factor
enhancing exercising. It is accepted that it is instead a compound of several intrinsic and
extra-musical factors stipulating motivation in the participant and therefore also the enhanced
outcome of the physical activity.
Previous findings in the field have been restricted to evaluating the ergogenics of music. In
recent years on the other hand, music’s motivational quality on a general level has become a
specific research topic.
As a result the Brunel Music Rating Inventory or BMRI (Karageorghis et al., 1999, 2006) was
designed: an instrument and conceptual framework evaluating the motivational qualities of
music 2 . In this inventory, motivational music (or ‘functional music’) is defined as “being
stimulating and inspiring physical activity […] simultaneously improving mood and reducing
perception of exertion” (2006, 899-900). Its opposite concept is oudeterous music which is
seen as music that neither motivates nor demotivates. The inventory therefore does not take
one but several musical factors as a starting point to define the motivational and enhancing
qualities of music for sports. The authors warn for an over-simplification of the term
motivational music, as it is believed that the separate musical properties influence the
motivation of the performer each in a different way by triggering either mood, attention and
arousal (2006, 906).
The first part of the BMRI validates the musical pieces on rhythmic qualities (tempo and
accentuation), harmonic and melodic aspects, cultural impact and personal association. These
factors are integrated in the BMRI via the use of 13 items or statements: familiarity, tempo
and beat, rhythm, lyrics related to physical activity, association of music with sport, chart
success, association of music with film or video, the artist, harmony, melody, stimulative
qualities of music, danceability and date of release (Karageorghis et al., 2006, 909). These
items are to be evaluated on a grading scale according to the tested piece of music.
2
The BMRI was designed by a group of researchers at the Brunel University and was first presented in a
research paper in 1999. Because of “limitations in its factor structure and its applicability to non-experts in
music selection” (Karageorghis et al., 1999, p. 899), the BMRI was redesigned and revalidated in a second paper
in 2006 (Karageorghis et al., 2006). Here the evaluated version of 2006 is discussed.
11
The outcome of this evaluation then comprises the motivational qualities of the certain song.
The result can be positive, which means the song is accepted as functional music for physical
exercising. The second part of the rating inventory assesses the relationship of the person with
the song in particular. Rhythm, style, melody (tune), tempo (speed), sound of the instruments
used and beat are questioned on being motivational to the performer. Based on the answers to
those six statements according to the music, an image is sketched of the motivational quality
with regard to the individual.
The inventory requires extensive administration when dealing with a collection of songs.
Moreover all statements and items of the two parts of the inventory carry the same weight in
the calculation of the motivational score of a music piece. To illustrate, the item ‘tempo of the
song’ carries the same importance as ‘date of release’ in calculating the functional quality of
music for exercising; or: the motivational effect of the rhythm is processed equally as the
motivational effect of the sound of the instruments used. Previous research has shown (Crust,
2004; Crust & Clough, 2006; Edworthy & Waring, 2006; Karageorghis et al., 1999, 2006)
that all intrusive factors in determining motivational quality of a musical piece differ in
agency. Consequently, the BMRI is in need for further investigation and refinement.
Despite these shortcomings, in light of the present study, the rating instrument is a significant
step towards a standardized method of music selection for physical exercising.
12
2.3 Towards a music selection system for physical exercising
This research was partly inspired by the idea to develop a system that would select music
from a given song collection for a sports exercise. Hereby in real time tracks would be chosen
wisely so that musical preference for the exercise, personal exercise habits, type of exercise
and in particular stage within the exercise would be respected.
One example of such a system is the Yamaha BODiBEAT, a device working as heart rate
monitor and music player, integrating the music playing as being depending on the heart rate
data. Song selection is based on its match with a predefined heart rate range, fixed throughout
the entire exercise duration. As a consequence of the music selection the exerciser exerts
himself less even when running intensity is held constant (Fukashiro, 2008).
A second already existing selection system is the IM4Sports (Interactive Music for Sports)
music system (Wijnalda et al., 2005); a system generating playlists on the basis of by the user
prescripted constraints for exercising, but also able to adapt or modify the playlist in realtime
whilst exercising where the situational performance quality demands. For instance in a
situation where heart rate pace has dropped from the target heart rate and needs to be restored.
Three steps characterize the system by Wijnalda et al. They incorporate teaching the system
the exerciser’s musical preference and exercise habits and defining the constraints of a
number of exercise schemas differing in duration of stages and intensity. Whilst running the
exercise and listening to the generated list of songs, the selection can be updated and
modified in real-time according to current heart rate and stride frequency (measured via steps
per minute) and its congruency with target heart rate and stride frequency – these last ones
predestined as constraints of the selected training scheme. Three strategies can perform the
modification: fixing the pace, matching the pace or influencing the pace.
In evaluating the two systems, we see that an ideal way of music selection for physical
exercising is within reach. Still, not all possible intrusive factors are dealt with. So does the
BODiBEAT perceive the physical exercise as one constant physical output in a fixed context.
It does not incorporate changes of intensity, consciousness and exertion occurring while
exercising.
13
The other example, the IM4Sports system was not tested empirically and hence misses proof
for its predicted positive effect on enjoyment and motivation for physical performance in
practice. Although user feedback belongs to the finalising of the generating and usage of the
playlist, motivational and affective response on the performance on the other hand does not.
This study extends the issue by investigating the possibilities with motivational music after it
has been selected in function of the sports performance on a level higher than the songs. Thus
it was designed to test the motivational qualities of the overall music organization; in other
words to investigate a way to let the order and timing of the right songs be right as well.
It strives at complementing the previously designed systems such as the BODiBEAT and the
IM4Sports with a human-oriented theoretical basis comprising all involved variables.
Furthermore the suggested theory will be tested in practice by an experiment.
14
2.4 Optimal Organization
On the basis of previous findings and literature the concept of Optimal Organization is
proposed. It suggests that next to a correct song choice, the selection of tracks has to be
organized optimally, so that likewise order and timing of the songs will affect the motivation
and performance quality of physical exercising 3 .
The concept consists of a collection or grammar of preconditions (for an overview see
attachment A) that helps the music ordering towards optimal organization. The preconditions
can be divided into three categories: internal musical factors, extra-musical factors and nonmusical factors. Internal musical factors are tempo, rhythmic strength (Crust & Clough, 2006;
Karageorghis & Priest, 2008), rhythmic pattern (Karageorghis & Priest, 2008), uplifting
melody and harmony (Karageorghis & Priest, 2008) and in sum these factors comprise into
motivational qualities of the music as defined by the BMRI.
Influencing factors related to the music (extra-musical factors) are: familiarity and musical
preference (North & Hargreaves, 2000; Crust, 2004; Karageorghis & Priest, 2008), musical
variety (DeNora, 2000; Karageorghis & Priest, 2008), positive or movement-related lyrics
(Karageorghis & Priest, 2008) and association with sport, exercise, triumph or overcoming
adversity (Karageorghis & Priest, 2008).
The rest of the factors do not have any direct relation to music (non-musical factors) and
consist of type of sports activity (Karageorghis & Priest, 2008), intensity of the activity
(Karageorghis & Priest, 2008), personality (Crust & Clough, 2006) and finally the exercise
stage which is the novelty in this research. Karageorghis and Priest (2008) have already
referred to the need of tailoring the music to various components of the training in the concept
segmentation phenomenon 4 , and DeNora (2000) has investigated the change of music in the
development of aerobic classes. These two sources are the basis of musical organization
according to exercise stages.
3
No knowledge can be obtained on what motivational music exactly evokes: does it affect the athlete’s
motivation, resulting in putting a higher effort into the workout or a less perceived exertion? Or does the music
affect directly the action of bodily movement? We know that the control of behaviour between stimulus and
response works on different levels (Toates, 2004) but with regard to the present topic no support was found.
4
So that for instance “work time and recovery time are punctuated by music that is alternatively fast and loud or
slow and soft. This approach is especially well-suited to highly structured sessions” (Karageorghis & Priest,
2008, 5).
15
DeNora (2000) has found in her observation of numerous aerobic classes that depending on
the song choice, the class becomes a more ordered environment resulting in a more efficient
exercise. Here, DeNora’s outcomes are taken as a framework for the theory of optimal
organization. It is extended with applications for individual exercising and complemented
with knowledge from other research.
An aerobic lesson is a highly controlled environment divisible in pre-scripted stages that
differ in type of movement and energy level. In the case of an aerobic lesson, distinction can
be made between warm-up, pre-core, core, cool-down and floor exercises, all requiring
differently selected music. The same idea is applied in optimal organization, but transferred
into warm-up, pre-core, core, ultimate end, cool-down and stretching. In order to keep up the
motivation, it is believed that the higher controlled the environment, the more efficient the
workout will be (DeNora, 2000). Music herein holds a leading position. For an overview of
the exercise in terms of its stages, see the template in figure 1.
Warm-up is the starting stage where participants have to get motivated and attracted to exit
everyday life and transit into the scenery of workout. Therefore catchy music is employed,
appealing the performers by lyrics, melodies, rhythms and/or orchestration. In particular lyrics
need to be positioned in the foreground. Starting songs, Karageorghis and Priest (2008) add,
need a tempo between 80-130 bpm, as for many the target workout tempo is 120 bpm.
Previous findings (Iwanaga, 1995; Karageorghis, Jones & Low, 2006) have indicated that
exercise heart rate and music tempo preference increase correspondingly. Therefore the ratio
‘current vs. target workout tempo’ will result in ‘current vs. target heart rate’. Music works as
a leader in facilitating this transition, in practice “successive tracks should create a gradual
rise in music tempo to match the intended gradual increase in heart rate” (Karageorghis&
Priest, 2008, 5). Especially during high-intensity workout (whereby the heart rate corresponds
with 75% of the maximal heart rate reserve) this relationship is considered to be highly
significant. Recent findings on the contrary have determined that for high-intensity workout
medium tempo (115-120 bpm) is preferred (Karageorghis, Jones & Stuart, 2008).
That the first song of a sports playlist has to be chosen accurately is additionally supported by
verbal feedback of participants of several studies (Tenenbaum et al., 2004), who reported that
the music at the beginning of the workout helped them a lot. This result foresees another
16
approval for the necessity of approaching music for workout with the exercise stages as basic
assumption.
Variety of songs is of utmost importance; DeNora and Karageorghis & Priest devote for
variation throughout the whole song selection in order to exclude monotony as much as
possible. It is said to be of key importance to get the sportsmen on track and to distract the
mind from any sensation of exertion.
Exercise
Intensity
A
B
C
D
E
Actual heart rate or actual intensity
Target heart rate or target intensity
Musical shift
A
B
C
D
E
F
warm-up
pre-core
core
ultimate end
cool-down
stretching
Figure 1. Training scheme with stages and actual/target heart rate
F
Time
17
After having launched the session a shift in the music occurs whereby lyrics diffuse towards
the background and eventually disappear while rhythm and beat take over the foreground.
Here starts the pre-core stage whereby the actual workout commences. DeNora points out the
necessity of a remarkable rhythmic pattern or “musical hop” here, providing a model for the
bodily action of the participant. Additionally, songs are required to follow-up logically so that
one leads to and fulfils into the next one. “These structures”, DeNora adds, “are achieved
through cadences and closures, through the generation and resolution of harmonic suspense,
through upward or cyclic melodic movement […] and through modulation upwards to higher
keys. Interlinked series of tonal and rhythmic structures in turn may be used to seduce bodies
to keep moving, to move on up or move on into the next level, the next unit” (DeNora, 2000,
98).
Keeping up the workout is facilitated by the power that music wields. During the core which
implies submaximal to maximal workout intensity, gear changes and the exercise becomes
more strenuous. Simultaneously the participants become less self-conscious and more
engaged in the workout. Based on the idea that tempo and beat are the central preconditions
for a correct selection and organization, the present work extends the use of the last model by
letting songs succeed according to an increasing number of beats per minute. This is supposed
to happen during the pre-core continuing throughout the core, until the maximum point of
workout.
At the ultimate workout intensity, a PowerSong (Nike + iPod, 2008) is exemplified to be
played. This type of song possesses the highest motivational quality of the song collection and
brings about positive effect to the participant. The PowerSong is labelled as the song giving
that extra boost to get the maximum out of the workout.
After the ultimate maximum energy output relief starts – which is the starting point of the
cooling down. Hereby music with a slow melody in the foreground moving on a background
in double time is suitable; very often well-known ballads are chosen. “Familiar melodies and
lyrics are brought back to the fore, and rhythm is relegated to the background” (DeNora,
2000, 101). Accordingly also songs with a decreasing number of beats per minute are
selected, in the line of a new – and lower – target tempo and heart rate. In this way the songs
yield the return to self-consciousness.
18
Hereafter stretching is being performed, accompanied by similar songs as played during the
cool-down, but possibly lowered in volume and lowered in pitch. Stretching or floor exercises
as in the case of aerobics ring in the closure of the entire workout.
We remember that in the overall scope of the exercise, two shifts occur – following the
direction of the target workout intensity – whereby one goes upwards and one downwards,
both divided by the peak of maximal workout intensity. The two shifts are also reflected in the
organization of the selected songs: from emphasis on melodic and harmonic aspects towards
temporal aspects in the first and vice versa for the second shift.
Optimal organization is a tool that can be used for any music for sports selection but is in the
scope of this research aimed at individually performed sports with a cyclic pattern such as
walking, running or cycling. Previous studies have pointed out differences in use of music and
professionalism of the athlete: the ergogenic ability of music seems to work in most of the
cases only for non-professional and semi-professional athletes. According to Brownley et al.
(1995) listening to music with a tempo higher than 160 bpm can be counterproductive in the
case of trained elite runners.
19
3. Experiment
3.1 Research questions and hypothesis
In this research the theory of optimal organization is tested empirically. An experiment is
carried out whereby participants will run alternately with optimally and non-optimally
asynchronously organized music as guidance. The results of this experiment will then show
whether the theory of optimal organization works, meaning that the performance quality is
enhanced; or not.
An answer is sought for the following research questions:
Does optimal organization of musical playlists enhance the performance quality?
How are the physiological and psychological responses of the performances related to each
other? Or in other words: does a higher motivation correspond with a better performance and
vice versa?
It is hypothesized that the results will show a better performance quality in the case of
listening to the optimal playlist.
This would mean a variance in:
-
physical condition via measuring heart rate
-
energy expenditure in spent calories
-
length of the exercise in time
-
psychological state or motivation via responses of exertion and general feeling of the
performance (in the text referred to also as affective response)
when comparing the two conditions.
The effect on the physical condition in the sense of heart rate is not clearly hypothesized, as
both increase and decrease could occur. Decrease can be expected due to a possible gaining of
confidence or achievement of flow. Yet, thinking back on the hypothesis of the length of the
exercise, we might expect an increase in heart rate for the optimal condition. Higher heart rate
and less performance time would then occur due to excitement or adrenaline as a response to
music as stimulus.
20
With regard to the length of the exercise, it is predicted that exercising with the optimal
playlist will take less time as a consequence of a higher motivation.
Concerning energy expenditure, as with heart rate, no straightforward prediction can be made
on how it will look like for optimal and random trials. No doubt on the contrary, according to
the theory a difference will be observed between the two.
Regarding the motivational responses it is predicted that rating of perceived exertion will be
lower in the case of listening to an optimal playlist; and the feedback on general performance
feeling will be better in the trials with the optimally organized music.
It is the aim of this exploratory research to help filling the lack of investigation in
motivational music for sports (Karageorghis & Terry, 1997) and to support progress in the
search for a wise music-selection system for physical exercising.
In addition, the research aims at finding general constraints in the outcomes of optimal
playlist organization use. The outcomes will therefore be screened for consistency within the
sample of the population initiating the construction of a model for optimal organization.
In order to test the research questions a highly ecological experiment was carried out.
21
3.2 Experimental design
3.2.1 Participants
A total of six participants were voluntarily recruited via the international and sports networks
of the University of Jyväskylä (Finland); all being between 21 and 39 years old (M = 28.2).
The gender division was respected so that three of the participants were female (M = 24.7
years old) and the other three male (M = 31.7 years old). All were semi-professional runners
running at least twice a week a distance longer than 5 km accompanied by music listening.
The participants all graded themselves as having a level of physical activity from 6 to 7.5
(whereby 0 = avoidance to no participation in physical activity and 10 = athlete on world class
level). They did not receive any reward for their participation and confidentiality was assured
throughout the experiment.
3.2.2 Procedure
3.2.2.1 Prior to trials
The participants filled out a questionnaire (see Attachment B) prior to the actual experiment
fathoming their exercise and music preference and habits. The questionnaires were returned
including 15 songs that they use or would intend to use for exercising. Each of these songs
was rated by likeness and familiarity of the particular song. Whether the song carried any
personal associations was additionally graded in the questionnaire.
Next, for each participant three playlists were designed with the person’s own songs that
scored equally in familiarity. Using their own music closes out the familiarity effect
(Karagorghis & Terry, 1997) and stands by the proof that all music genres can be used 5 . Two
playlists were allocated randomly whereas one song list was designed with an optimal
organization according to the theory. This implies that all tracks were analyzed on their
motivational qualities, namely tempo in bpm, rhythmic characteristics, melodic clarity, lyrics
5
Most music chosen by the participants for the experiment and intended for exercising tends to be pop, techno or
rock music. This leads to two possible explanations: either these genres are seen as the most eligible for sport
activities or either the selection of participants was not diverse enough.
22
and association and internal variety amongst each other. Songs under 90 bpm without doubletiming were excluded from the selection or added in the cool-down section.
The total duration of all the songs together was around 35 minutes. It was estimated it would
take in total 30 minutes for the participants to accomplish the track, therefore roughly the first
30 minutes are the songs of the actual playlist, completed with an extra song extending the
last sensation if necessary.
After distributing the songs in the division warm-up, pre-core, core, highest intensity and
cool-down, the songs were ordered. This was done with an increasing tempo, starting with
high (female) voices; increasing the rhythmical emphasis and energetic quality; keeping the
variation of genre and instrumentation and ending with a kind of powersong followed by a
lighter track that prefaces soothing the exhaustion. At last also the description of the personal
feelings and associations of the participants were analysed and taken into consideration for the
design of the playlists 6 .
6
A consequence of having a limited number of preferred songs by each participant is that some sets are more
eligible for creating a workout playlist than others. Some selections for instance did not contain any good warmup songs, not enough variation in tempo or consisted of very long songs.
23
3.2.2.2 Building the optimal playlist: two examples
Tables 1 and 2 below show the optimal playlists designed for the experiment of respectively
participant 1 and 6. The listed songs are accompanied by their tempo (in bpm), track duration,
total exercise duration, relevant important musical descriptors and feelings or associations of
the participant as stated in the preceding questionnaire.
Track
Nr.
Tempo
(in
bpm)
Track
Duration
Total
Exercise
Time
1
127
03:43
03:43
2
125
04:27
08:10
Musical
descriptors
Personal feelings or
associations
Female voice
Lyrics
High melody
"This song gives me good mood
and energy, nice music with sensual
lyrics."
"A repetitive but awesome rhythm
either accompanied with a powerful
bass or a soft melody. It probably
doesn’t work on everybody, but I’m
totally receptive to that kind of
sounds, I love it."
3
126
07:11
15:21
Rhythmical
Repetitive
4
130
03:20
18:41
Rhythmical
5
139
06:38
25:29
6
140
02:07
27:36
7
140
05:44
33:20
8
128
06:27
39:47
Powersong
"It boosts me to go faster or keep
motivation."
"Very aggressive music. It is just
brutal techno. I would get a heart
attack if listening to that for the
whole training session. Can’t help
moving when I hear it"
Motivating
lyrics
Table 1. Optimal playlist of participant 1.
The music selection of participant 1 belonged to the musical genre techno. Rhythm stands
hereby more in the foreground whereas melody and lyrics are of a relatively minor
importance. As a consequence the list was suitable for endurance maintenance but less
suitable for warm-up. In general the songs are long and do not vary much amongst each other.
The descriptions by the participant partly compensated these difficulties; so that the ordering
could be administered conform to the theory. Increase in the tempo was achieved but the
range (127-140 bpm) rather small compared to the other music lists.
24
Track
Nr.
Tempo
(in
bpm)
Track
Duration
Total
Exercise
Time
1
124
03:57
03:57
2
108
04:30
08:27
3
118
04:02
12:29
4
118
04:26
16:55
Rhythmical
Motivational
lyrics
“Groove, boosting”
5
140
04:24
21:19
Melodic
“Fun, energy, inspiration”
6
157
04:13
25:32
Powersong
“Electric, energetic,
motivation”
7
200
03:23
28:55
8
115
03:04
31:59
9
86
06:02
38:01
Musical
descriptors
Female voice
High melody
Bright melody
Female voice
Rhythmical
Personal feelings or
associations
“Upbeating”
Motivational
lyrics
High melody
Table 2. Optimal playlist of participant 6.
Participant 6 chose for pop music from the eighties and songs in the current pop and rock
music charts. Genre, tempo (range 108-200 bpm) and rhythm varied sufficiently and the
largest part of the songs were sung by a high female voice. What restrained the selection from
making a suitable playlist were the big leaps in tempo amongst the songs. The rise in tempo
did therefore not result as smoothly.
3.2.2.3 Execution of trials
All participants completed four running exercises while listening to one of the designed
playlists (two random playlists - R1 and R2; and twice the optimal playlist - O1 and O2) and
wore a device tracking physiological responses heart beat and energy expenditure in Kcal.
The device also calculated total length of the exercise in time. The order of conditions
exposed differed for all participants and each test was performed individually. The playlist
was meant to work as asynchronous music so any synchronization of bodily movement with
the tempo of the music occurs without conscious effort of the participant.
25
After each exercise verbal feedback was given by the participant on rate of exhaustion, the
participant’s general feeling of the performance and any failures due to their own exercise
habits or the experiment’s setup 7 (see Attachment C). The rate of exhaustion was collected via
the Borg-scale for Rating of Perceived Exertion (RPE) (see 3.2.3 instruments). The Borgscale represents a subjective feedback of the exercise and has been previously used
successfully in music and sports researches and settings (Boutcher & Trenske, 1990;
Fukashiro, 2008). Fukashiro has suggested that the RPE depends on the music that is listened
to regardless differences in pace (2008, 5). General feeling of the performance was rated on a
rating scale.
Further in this work RPE and general feeling will be referred to as the psychological or
affective responses. Although the designation ‘psychological’ is not completely valid in its
sense, here it is chosen as it approximates more to motivational significance.
Considering the participant’s own exercise habits, all performed the exercises outside on the
same route with a distance of 4500 m. The track did not have differences in altitude, was free
from motor vehicles and was surrounded by nature. No extreme differences in weather
conditions amongst the exercises were observed.
The four exercises were performed on non-consecutive days (with at least an interval of two
days in between) and on the same time of the day (morning – noon – afternoon – evening).
The participants had the voice to refuse or cease performing at any time during the
experiment’s schedule. Only once a participant chose to discontinue and suspend an exercise
after completing the trial partially due to abdominal pain. This exercise was retaken later on
within the same conditions.
7
The feedback on possible failures is initially not planned to be processed in the analysis and results of the
research. If the reports contain information that could confound the outcomes of the research on the contrary,
they will be adopted afterwards.
26
3.2.3 Instruments
Making the playlists 8 .
For the tempo analysis of the tracks the BPM Analyzer MixMeister (2008) was used.
Recording physiological respones.
The runners wore a Suunto T6 heart rate tracking device which consists of a heart rate belt
and watch. All participants were familiar with the use of a training device and apart from one
fall off of the belt – where the exercise was retaken later on with the same conditions, no
anomalies happened with the Suunto device.
From the T6 tracked physiological variables only log time, heart rate (with a sample size of
ten seconds) and spent calories were used 9 . The data was downloaded and exported for
statistical data analysis to SPSS with the latest update of the Suunto Training Manager,
developed by Firstbeat Technologies.
Music was played by portable mp3-player SanDisk Sansa c240 and listened to through the
participant’s own headphones or earplugs; loudness was freely set according to personal
preference.
Recording psychological responses.
The feedback on perceived exertion was collected using the Rating of Perceived Exertion
(RPE) scale by Borg (1989) whereby participants rate their feeling of exertion (or exhaustion)
on a scale from 6 to 20 (whereby 6 = “no exertion at all” and 20 = “maximal exertion”). These
digits represent and relate with the heart rate extremes of an adult, namely 60 bpm (low
resting HR estimate) and 200 bpm (1989, 30).
Collecting feedback on general feeling of the performance was administered using a Likert
scale from -2, -1, 0, +1 to +2 with the items ‘bad’, ‘could be better’, ‘average’, ‘good’ and
‘excellent’ assigned respectively. It was told to the participants to rate the overall general
8
In the song selection and ordering the Brunel Music Rating Inventory (Karageorghis et al., 1999, 2006) was not
used due to the equality of factors (see 2.2). However, all these factors were incorporated separately in the
making of the playlist as described under 3.2.2.2.
9
Also variables oxygen consumption, EPOC, respiration rate, ventilation and energy consumption were tracked
with the Suunto T6. The reason these variables were not used for this research is their dependence on the heart
rate variability and therefore not adding new data.
27
feeling they had when performing the exercise, not excluding the current state at the moment
they filled out the questionnaire.
The opportunity was given to report on failures during the exercise, categorized as being due
to the experiment’s setup or due to personal exercise habits.
All these data were collected orally and directly written down by on a form that can be found
as Attachment C.
28
4. Analysis and results
4.1 Data analysis
Statistical analysis of the data was done using SPSS (version 17.0). The heart rate data were
screened for outliers and missing values were imputed using linear interpolation followed by
smoothing using running medians with a five-point span. All data met the normality
distribution.
A one-way repeated measures analysis of variance (ANOVA) was performed to determine the
increasing or decreasing effect of optimal organization on the dependent variable heart rate.
This was done for the four trials of each participant. In order to determine differences between
optimal and random heart rate responses, t-tests were performed likewise on all participants
individually. Statistical tests and comparisons amongst participants were not performed since
heart rate range and physical activity level (as reported in the preceding questionnaire)
differences were substantial and not suitable for mutual comparison.
For the other physiological variables energy expenditure and length of the exercise in time
means of optimal and random trials were calculated and compared.
Exploring the rating of perceived exertion, means were calculated and compared for general
trends amongst all participants. Afterwards differences were tested statistically via analysis of
variance between the four trials and a paired-samples t-test between the two optimal and the
two random conditions. Repeated measures analysis of variance and dependent samples t-test
was likewise performed for the variable general performance feeling.
In addition attention was paid to the behaviour of participants cross-variably via comparing
the responses and results of all dependent variables. As a consequence a predictive model is
proposed estimating physiological and psychological behaviour as response to optimal
playlist impulse.
29
4.2 Physiological responses: heart rate data
At first an ANOVA was executed in the heart rate data to determine any existing difference
between the four trials for each participant. The results as shown in table 3 reveal that in 5 of
the 6 participants there exists a significant difference between the four trials’ mean values.
One participant’s data did not indicate significance. A main difference was found for
participant 1 (F3,520 = 320.024, P < 0.001).
Participant
Mean
SD
O1
O2
R1
R2
149.06
141.47
136.71
142.08
3.348
3.781
2.560
3.396
O1
O2
R1
R2
146.12
142.87
140.43
143.29
4.484
11.650
3.513
6.019
O1
O2
R1
R2
171.59
170.81
173.11
156.98
5.842
6.642
5.813
9.673
O1
O2
R1
R2
158.23
157.02
157.46
160.54
7.635
5.378
9.461
12.837
O1
O2
R1
R2
168.97
163.44
167.20
163.19
3.960
4.210
4.189
3.290
O1
O2
R1
R2
169.13
170.30
166.92
168.61
7.159
11.879
7.246
7.883
F
df1, df2
320.024*
3, 520
24.116*
3, 727
30.612*
3, 655
1.029
3, 663
57.219*
3, 606
25.439*
3, 703
1
2
3
4
5
6
* P < 0.001
Table 3. Repeated measures ANOVA for all participants testing differences between the four
performance trials O1, O2, R1, R2.
30
Although the analysis of variance showed a significant difference between the groups’ means,
Tukey HSD post hoc tests did not show consistency in the hypothesized relationship within
the random trials and the optimal trials. Notwithstanding the lack of significance, a stronger
relationship between the two optimal trials is noticed for all participants.
In order to clarify the tendency of the optimal trials found via the ANOVA, a related t-test
was performed to test the difference between optimal and random trials for all participants.
Participant
Difference of
the mean
SD
t
df
1
6.085
5.164
18.816 *
254
2
1.045
3.961
4.984 *
356
3
1.077
3.626
5.354 *
324
4
1.240
7.733
2.881 **
322
5
-0.023
6.029
-0.067
300
6
-1.741
10.222
-3.168 **
345
* P < 0.001
** P < 0.01
Table 4. Related t-test results for all participants testing the differences between optimal and random
playlist conditions.
For 5 out of 6 participants optimal and random heart rate data were significantly different, of
which three whereby P < 0.001. A main difference was found by participant 1 (t254 = 18.816,
P < 0.001). Participant 6 showed a reverse significance. Participant number 5 did not show
any difference in heart rate on the optimal playlist.
The hypothesis of this research presupposed a lower mean heart rate in the condition of the
optimal playlist. If we take a look at the means of the four trials in table 3, we can derive an
opposite result. This means the empirical testing has shown that optimal playlist listening
during exercising excites and therefore lets heart rate increase.
Thus, for four participants the optimal playlist facilitated a significantly higher heart rate. For
participant six, a negative relationship is recorded (t345 = -3.168) pointing at a significantly
lower heart rate in the optimal condition. Whereas this last one, participant 6, has means for
31
the four trials that correspond with the previous outcome of the analysis of variance (being
O1: 169.13; O2: 170.30; R1: 166.92; R2: 168.61), the large standard deviation of the second
optimal trial can clarify the negative difference of the mean between total optimal and random
trials (for the t-test) and consequently also the smaller within difference of the optimal cases
than the randomly allocated cases.
Recalling the results of the precedingly performed analysis of variance suggesting a
consistency between the two optimal trials’ results, we can add and conclude that the optimal
exercises do indeed differ significantly from the randomly allocated ones for five out of six
participants.
32
4.3 Spent calories
Next to heart rate also the expenditure of calories was measured during each performance. If
we look at the total expenditure of each trial for each participant (visualized in table 3), we
see that apart from two cases (participants 1 and 2) all random trials were characterized by a
higher expenditure of calories as compared to the optimal condition. This does not confirm
the hypothesis that there was to be seen a difference between the two trial conditions. The
tendency that optimal playlists reduce the energy expenditure could embody a clearer answer
to the other hypothesis at the beginning of this research that no predictions can be made
whether optimal or random will rule over in amount of energy expenditure.
Participant
Optimal 1
Optimal 2
Random 1
Random 2
Optimal
Random
1
348
288
298
285
318
291.5
2
319
300
300
309
309.5
304.5
3
341
346
342
345
343.5
343.5
4
470
488
487
500
479
493.5
5
331
340
342
345
335.5
343.5
6
380
366
377
376
373
376.5
364.833
354.667
357.667
360
359.75
358.835
Mean
Table 5. Energy expenditure for optimal and random trials (in Kcal).
A one-way repeated measures analysis of variance has however shown no significant
difference between the four conditions (F3,20 = 0.024, P > 0.05), neither did a repeated
measures t-test between all random and optimal conditions (t11 = 0.175, P > 0.05). A more
elaborate number of subjects could bring more clarity in the significance testing.
33
4.4 Length of the exercise
The third variable connected to the physiology of the exercise is the total length in time of the
exercise. Table 4 below contains the total exercise times for the four trials in seconds. The last
two columns represent the means of the optimal and random exercises.
We can derive from the grid that in 4 out of 6 participants the randomly exposed exercises
lasted in total longer than the optimal ones. One participant performed the same average
duration for optimal and random conditions and one participant performed longer in the
optimal than in the random conditions.
Comparing optimal and random condition time values statistically with a t-test did not reveal
any significance (t11 = -0.713, P > 0.05).
Participant
Optimal 1 Optimal 2
Random 1
Random 2
Optimal
Random
1
1400
1310
1358
1416
1355
1387
2
1850
1841
1920
1840
1845.5
1880
3
1690
1683
1640
1733
1686.5
1686.5
4
1638
1711
1703
1767
1674.5
1735
5
1510
1560
1520
1623
1535
1571.5
6
1815
1865
1813
1714
1840
1763.5
1650.5
1661.666
1695
1682.166
1656.083
1673.583
Mean
Table 6. Exercise durations (in seconds).
34
4.5 Affective responses: perceived exertion and general feeling
4.5.1 RPE
After each performed trial the participants were asked to estimate their exertion on the RPE
Borg-scale. Table 3 illustrates the results for each participant’s optimal and random trials,
their sum and all means.
Participant
Optimal 1 Optimal 2
Random 1
Random 2
Optimal
Random
1
13.00
14.00
14.00
11.00
13.50
12.50
2
11.00
13.00
13.00
14.00
12.00
13.50
3
13.00
11.00
13.00
13.00
12.00
13.00
4
10.50
11.00
12.00
11.50
10.75
11.75
5
11.00
15.00
12.00
15.00
13.00
13.50
6
13.00
13.00
12.00
14.00
13.00
13.00
11.9167
12.8333
12.6667
13.0833
12.375
12.875
Mean
Table 7. RPE data for the six participants’ optimal and random trials including the means
of the two conditions.
One out of six participants (participant 1) reported a higher exertion after listening to the
optimal playlist. One has reported in average equally to both optimal and random trials
(participant 6). All other participants assigned more exertion to the randomly allocated
playlist, of which three reported a difference of 2 between random and optimal. In other
words; four of the six participants have perceived the optimal playlist as less exhausting than
the randomly allocated one.
Again for rating of perceived exertion a paired-samples t-test for optimal and random
conditions did not show any significance (t11 = -1.239, P > 0.05), neither did a repeated
measures analysis of variance for all four trials (F3,20 = 0.851, P > 0.05).
35
4.5.2 General feeling
Next to RPE participants evaluated and reported their general feeling of the performance. For
this five possible answers were given: ‘bad’ = -2, ‘could be better’ = -1, ‘average’ = 0, ‘good’
= 1 and ‘excellent’ = 2.
Participant
Optimal 1
Optimal 2
Random 1
Random 2
Optimal
Random
1
-1
0
0
-1
-0.5
-0.5
2
0
1
1
0
0.5
0.5
3
-1
0
-1
-1
-0.5
-1
4
-1
-1
-1
-1
-1
-1
5
1
-1
0
-1
0
-0.5
6
2
0
1
-1
1
0
Mean
0
-0.1667
0
-0.8333
-0.5
-0.4166
Table 8. General feeling data for the six participants’ optimal and random trials
including the total optimal and total random responses.
We can derive from table 4 that for all participants the general feeling of the performance was
better or equal than in the case of the optimal playlist10 . Two participants (number 1 and 2)
have reported similarly in the two conditions throughout the four trials: therefore the means of
optimal and random conditions are equal.
The means however of the optimal and random exercises together show another direction of
results: here the general feeling of the optimal condition was not as good as the random. In
general we can say that the differences are small and negligible.
A dependent samples t-test for the optimal and random conditions did not indicate significant
differences (t11 = 1.483, P > 0.05) and also a repeated measures analysis of variance did not
recognize a difference between the 4 trials (F3,20 = 1.206, P > 0.05).
10
Participant 4 added to his rating afterwards that he cannot feel a performance as being good or excellent. For
him there are always improvements to make. Therefore he answered to the question what his general feeling was
of the performance in all cases ‘could be better’.
36
4.5.3 RPE vs. general feeling
Comparing the results of exertion and general feeling from the above, it is expected to have a
negative linear relationship between the two variables since exhaustion and general
performance feeling belong to the given moment’s affective or psychological state. The
feeling one has in general of a performance is therefore related to the extent to which one is
exhausted from it. Statistically we can speak only of a tendency towards a negative linear
relationship or correlation. A bigger sample size could be more decisive in supporting the
negative linear relationship.
37
4.5.4 Mean heart rate vs. RPE
The graphs (Figure 2) below indicate the mean heart rate and perceived exertion for each of
the four exercises of each participant separately. However these data should not be compared
between but within participants, we can see that the heart rate range amongst the participants
differs. Participant 1 has a lower range but the most diverging extremes, participant 6 the
highest range and participant 4 does not have any range at all: all values of mean heart rates
for all exercises almost do not fluctuate at all. For almost all of the trials the participants rated
their perceived exertion as being under their mean heart rate value 11 apart from two cases of
participant one.
One of our research questions was whether physiological and psychological-motivational
responses of the performances were related to each other. And if so: how?
It was predicted that both would go hand in hand, so that when listening to an optimal playlist,
the mental state would go up and this would then result in a better performance. Whether this
would mean a higher mean heart rate: namely that the participant’s motivation would make
him or her run faster; or a lower heart rate meaning that a positive mental state would result
physiologically in a better performance comfort and therefore no increase of heart rate, is not
clear. Both explanations are feasible so the experiment’s data will have to rule out.
Below we can see the mean heart rate for the four exercises in function of its rating of
perceived exertion for all participants. It was predicted that the first two bars of the mean
heart rate would differ in height with the last two bars (no directional outcome was
hypothesized whether heart rate would increase or decrease in optimal organization setting).
Regarding the rating of perceived exertion, it was predicted that the first two ratings would be
(in average) lower than the last two random ones.
Participant 5 shows support for the secondly proposed hypothesis: a lower mean heart rate
goes with a higher perceived exertion. On the contrary these data are not significant in the
context of optimal and random conditions. Participant 1 shows a trend towards support for the
11
The RPE-scale is designed in that way that the item values correspond with the adult heart rate values. An
RPE of 10 therefore corresponds with a heart rate of 100 bpm. Reliability tests on parallelism with HR showed a
high correlation above 0.90 (Borg, 1989, 31).
38
first hypothesis. Apart from the first optimal trial, the ratings of perceived exertion are in
accord with the heart rate data.
The graphs of the other participants do not show any consistency towards one of the
hypothesis.
Previously an urge was incited on a clearly higher rating of perceived exertion in the
randomly allocated exercise settings. Here however not the mean RPE values were employed.
Therefore no valid significant outcome can be derived from these comparisons. No definite
answer can be formulated on the research question whether physiological and psychological
responses relate to each other.
39
Figure 2.Mean Heart Rate and RPE for the exercises Optimal 1, Optimal 2, Random 1
and Random 2.
40
4.6 Modelling optimal playlist behaviour
The purpose of this cross-section analysis whereby each individual participant is taken as a
point of view throughout all the results and analysis is to explore and discover pattern(s) of
behaviour. When a clear pattern of behaviour in all dependent variables exists, it will be
possible to build a predictive model with which behaviour of the dependent variables can be
estimated. The outcome model can help the development of a music-selection system for
physical exercising as introduced in chapter 2.3.
In the following table, the results of each participant on the dependent variables are classified.
Items in bold indicate an anomaly in comparison with the other items of that variable.
HR
ANOVA: sig. P < 0.001
Participant 1
t-test: sig. P < 0.001
Optimal higher
ANOVA: sig. P < 0.001
Participant 2
t-test: sig. P < 0.001
Optimal higher
Energy
Optimal
less
Optimal
Length
Optimal
shorter
RPE
Optimal
higher
more
Optimal
shorter
Optimal
lower
Equal
Equal
Optimal
lower
Optimal
less
Optimal
shorter
Optimal
lower
Optimal
less
Optimal
shorter
Optimal
lower
ANOVA: sig. P < 0.001
Participant 3
t-test: sig. P < 0.001
Optimal higher
General
Feeling
Equal
Equal
Optimal
better
ANOVA: not sig.
Participant 4
t-test: sig. P < 0.01
Optimal higher
ANOVA: sig. P < 0.001
Participant 5
t-test: not sig.
ANOVA: sig. P < 0.001
Participant 6
t-test: neg. sig. P < 0.01
Optimal lower
Optimal
less
Optimal
longer
Table 9. Participants’ dependent variables overview
Equal
Optimal
better
Optimal
Equal
better
41
Apart from minor deviations 12 (indicated in bold) in the variables energy expenditure, length
of the exercise and RPE most of the responses are observed as either conform to the general
trend of the variable (optimal being either higher or lower for the variable than the random
condition) or the sum of both optimal and random values are equal. Concerning general
performance feeling, 50 % of the participant did not perceive any difference and the other half
found the feeling enhanced in the optimal playlist condition.
Note that apart from one (participant 6), none of the participants deviates in more than one
items of the general trend. Participant 3 tends to perform rather equally in random and optimal
conditions for the psychological variables13 and energy expenditure and exercise duration. Its
heart rate data however showed both in the ANOVA (F = 30.612, p < 0.001) and t-test (t =
5.354, p < 0.001) a high significance. Also participant 4’s results are relatively conform to the
general trend.
If we do not include the minor abnormalities and consider them as errors, we can extract
following model from the outcomes.
12
These minor abnormalities seem to be random and not depending on a consequently different behaviour of the
participant. Also errors due to material and setup of the experiment could turn out as cause.
13
For participant 3 were the RPE values O: 12 and R: 13 and the rating of general performance feeling O: -0.5
and R: -1.
42
IMPULSE
optimal playlist
as compared to a randomly
allocated playlist
RESPONSE
physiological
Higher heart rate
Lower energy
expenditure
Shorter Exercise
time
psychological
Lower perceived
exertion
In 50% of the
participants a better
general feeling
Figure 3. Predictive model for optimal playlist use in sports setting.
The model plots the mechanism of optimal playlist use in an exercise setting based on a
stimulus-response relationship. Using a playlist for physical activity will trigger physiological
and psychological responses; in other words affecting both body and mind. When the song
selection and ordering are done using optimal organization, these responses will act optimal
accordingly. This means for physiological responses that heart rate will increase, spent Kcal
will go down and less time will be needed for exercising. Psychological effects are a lower
perceived exertion and a 1 out of 2 chance for having a better general feeling. Randomly
allocated playlists do not bring about these responses to this extent 14 .
14
Random song collections are assumed to enhance performance quality in contrast to no music conditions
(Crust, 2004)
43
5. Conclusion
Earlier incite was given on the points of limitation in this research. In the following these will
be clarified and extended with implications for future research directions.
5.1 Limitations of the experimental design
That more confusion than clarification arose in the results of this study was partly owe to a
lack of quality in the design of experiment.
Participants
The small sample size is certainly a large shortcoming of this research. More participants
would have introduced more accurate results and a more truly representation of the
population. Although recruitment of the six participants went throughout one network, their
background, physiology and exercise habits differed 15 . The data showed a different heart rate
range and shape and also the length of a usual exercise was not in accord. Working for
instance with the participant’s own exercise intensity values instead of taking the same
intensity for all as done here, could balance out some of the internal differences. Balancing
between the search for similarity among the sample group and opening up the population
represented remains a challenge.
Making the playlist
Incongruences in personal background popped up throughout the music selections and their
differences in genre preference. In the cases where the researcher was not familiar with the
musical genre 16 , the issue arose that a lack of knowledge obstructed the playlist design.
In addition we have seen that every collection of songs is not as suitable to make an optimal
playlist from.
15
Two of the participants for instance were semi-marathon runners therefore not familiar with an exercise
running distance of 4.5 km.
16
The composition of participant 1’s playlists is an example of this issue.
44
A better knowledge of the participant’s relationship to the songs is needed. In this way the
playlist will become more individualized and contextualized and for instance more musical
surprises to keep the exerciser on track could be incorporated into the playlist. A
complementary interview after the questionnaire could offer solutions.
Data collection
A limitation in the design of the study was the device used for recording the heart rate. Its
accuracy of 10 sec was an imperfection due to which a lot of data has been missed out. Using
newer devices with more memory capacity and a finer recording resolution are a must for
similar experimental designs.
When starting a trial, participants had to start heart rate tracking device and music player at
the exact same moment. Since this is physically almost impossible also in future research it is
suggested to find a way to optimize the synchronization between heart rate device and music
player.
Collecting affective data after the performance was done through the RPE-scale of Borg and a
5 point-scale where participants rated their general feeling of the performance. This last one
has shown that small and almost negligible differences exist between the general feeling of an
optimal condition and a randomly allocated one. This lack of clear outcomes can be solved
via the use of a more-point Likert scale with a larger range so that nuances can be reported
more easily.
In addition, investigation is needed in the psychological-affective response recordings. The
present research has neglected the psychological state prior to the performance. Although the
state after the performance has been well-observed in this case, no knowledge is gained in the
possible positive or negative change that music and sports have facilitated.
45
5.2 Towards a music selection system for physical exercising
The outcomes of this research conveyed new insight in how structuring music along physical
workouts can be done efficiently in that the performance benefits from the musical input. It
pointed out that not only the correct selection of music, but also the ordering of that selection
plays a crucial role when optimizing a physical performance throughout music.
Moreover the presented optimal playlist use model contains the theoretical base for the
development of an intelligent music selection system as explained in 2.3. Thinking back on
the stimulus-response relationship of the model, we have seen that music initiates
psychological and physiological outcomes. Optimal organization of music will bring about
these outcomes in an enhanced manner.
The current research’s experiment and the outcome model put first the (optimal) music as
independent and then the outcomes as dependent variables. The wise music selection system
on the contrary has music as being dependent on the physiological and psychological
outcomes. Therefore, such a system can start to be built from the inversed model, so that
music is in other words the response of body and mind as stimulus.
The starting points of the IM4Sports (Wijnalda et al., 2005) and BODiBEAT (Fukashiro,
2008) were comparable and succeeded in developing an intelligent music selectioning system.
This research has justified and retested the assumptions of Wijnalda and colleagues and
Fukashiro theoretically. In addition, this study has provided a light from the musical science’s
point of view and contains the potential to develop such a music structuring system for
physical activity towards a humanly oriented system. This means taking into account the
musical, personal and environmental context of the exercise. A device with the ability to
moreover select the correct music along the physiological need of the exerciser and the
exercise’s stages could aid physical exercising, and is therefore the next step towards an
intelligent music organizing system.
46
5.3 Enhancement of the performance quality?
Conversely, the question still remains: do the psychological and physiological outcomes as
sketched in the model under 4.7 enhance performance quality? We have seen – roughly
speaking – an increase in heart rate, a shortening in time, a reduction of spent calories, a
descent in perceived exhaustion and in half of the cases a better general performance feeling
as consequences of using optimal playlist organization. But are these the desired results of a
training or competition? An elite athlete willing to win a competition with full intensity
probably does, but an athlete doing endurance training perhaps not. Besides it is the role of
athlete and coach to build this knowledge into training strategies and schedules. Additionally,
can the collection of physiological responses be extended with endurance-related and stride
frequency effects? Is optimal organization a tool for merely short intensive training while
facilitating a good mood?
More questions remain and further investigation is called on. If we recall the use of music in
sports settings we remember that mostly non-elite sports people doing training employ more.
Further investigation on different needs in performance and training settings and
corresponding musical constraints can bring about clarification and intensification of the
matter.
47
References
Acevedo, E. O. & Ekkekakis, P. (Eds.) (2006). Psychobiology of Physical Activity.
Champaign, IL: Human Kinetics.
Bishop, Daniel T.; Karageorghis, Costas I. & Liousou, Georgios. (2007). A Grounded Theory
of Young Tennis Players' Use of Music to Manipulate Emotional State. Journal of Sport
and Exercise Psychology, 29, 584-607.
Borg, G. (1989). Borg’s Perceived Exertion and Pain Scales. Champaign, IL: Human
Kinetics.
Boutcher, S. H. & Trenske, M. (1990). The effects of sensory deprivation and music on
perceived exertion and affect during exercise. Journal of Sport and Exercise Psychology,
(12), 167-176.
Brownley, K. A.; McMurray, R. G. & Hackney, A. C. (1995). Effects of music on
physiological and affective responses to graded treadmill exercise in trained and untrained
runners. International Journal of Psychophysiology, 19, 193-201.
Bull, M. (2000). Sounding out the city: Personal stereos & management of everyday life.
Oxford: Berg Publisher.
Crust, L. (2004a). Carry-over effects of music in an isometric muscular endurance task.
Perceptual and Motor Skills, 98, 985-991.
Crust, L. (2004b). Effects of familiar and unfamiliar asynchronous music on treadmill
walking endurance. Perceptual and Motor Skills, 99, 361-368.
Crust, L. & Clough, P. J. (2006). The influence of rhythm and personality in the endurance
response to motivational asynchronous music. Journal of Sport Sciences, 24(2), 187-195.
Csíkszentmihályi, M. (1990). Flow: the psychology of optimal experience. NY: Harper &
Row.
DeNora, Tia. (2000). Music in Everyday Life. Cambridge: CUP.
48
Edworthy, J. & Waring, H. (2006). The effects of music tempo and loudness level on
treadmill exercise. Ergonomics, 49(15), 1597-1610.
Fukashiro, S. Verification of effects on reducing feelings of fatigue during exercise using the
yamaha bodibeat. Retrieved January, 31, 2008, from
http://www.yamaha.com/bodibeat/FORMS/Yamaha_Univ.pdf.
Hanin, Y.L. (2007). Emotions in Sport. Current Issues and Perspectives. In: Tenenbaum, G. &
Eklund, R.C. Handbook of Sport Psychology (3rd ed). Hoboken, N.J., Wiley.
Hargreaves, D. J. & North, A. C. (1999). The functions of music in everyday life: Redefining
the social in music psychology. Psychology of Music, 27, 71-83.
Iwanaga, M. (1995). Harmonic relationship between preferred tempi and heart rate.
Perceptual and Motor Skills, 81, 67-71.
Karageorghis, C. I. (1999). Music in sport and exercise: Theory and practice. The Sport
Journal, 2(2).
Karageorghis, C. I. (2008). Music in sport and exercise: An Update on Research and
Application. The Sport Journal, 11(3).
Karageorghis, C. I.; Jones, L. & Low, D. C. (2006). Relationship between exercise heart rate
and music tempo preference. Research Quarterly for Exercise and Sport, 77(2), 240-250.
Karageorghis, C. I.; Jones, L. & Stuart, D. P. (2008) Psychological Effects of Music Tempi
during Exercise. International Journal of Sports Medicine, 29, 613-19.
Karageorghis, C. I.; Priest, D.; Terry, P. C.; Chatzisarantis, Niklos L. D. & Lane, A. M.
(2006). Redesign and initial validation of an instrument to assess the motivational
qualities of music in exercise: The Brunel music rating inventory-2. Journal of Sport
Sciences, 24(8), 899-909.
Karageorghis, C. I.; Mouzourides, D. A.; Priest, D.; Sasso, T. A.; Morish, D. J. & Walley, C.
L. (2009). Psychophysical and Ergogenic Effects of Synchronous Music During
Treadmill. Journal of Sport and Exercise Psychology, 31, 18-36.
49
Karageorghis, C. I. & Terry, P. C. (1997). The psychophysical effects of music in sport and
exercise: A review. Journal of Sport Behaviour, 20, 54-68.
Karageorghis, C. I.; Terry, P. C. & Lane, A. M. (1999). Development and initial validation of
an instrument to assess the motivational qualities of music in exercise and sport: The
Brunel music rating inventory. Journal of Sport Sciences, 17, 713-724.
Macone, D.; Baldari, C.; Zelli, A. & Guidetti, L. (2006). Music and physical activity in
psychological well-being. Perceptual and Motor Skills, 103, 285-295.
North, A. C. & Hargreaves, D. J. (2000). Musical preferences during and after relaxation and
exercise. The American Journal of Psychology, 113(1), 43-67.
Simpson, S. D. & Karageorghis, C. I. (2006). The effects of synchronous music on 400-m
sprint performance. Journal of Sport Sciences, 24(10), 1095-1102.
Szabo, A.; Small, A. & Leigh, M. (1999). The effects of slow- and fast-rhythm classical
music on progressive cycling to voluntary physical exhaustion. Journal of Sports
Medicine and Physical Fitness, 39, 220-225.
Tenenbaum, G.; Lidor, R.; Lavyan, N.; Morrow, K.; Tonnel, S.; Gershgoren, A.; Meis, J. &
Johnson, M. (2004). The effect of Music type on running perseverance and coming wit
heffort sensations. Psychology of Sport and Exercise, 5, 89-109.
Toates, F. (2004). Cognition, motivation, emotion and action: a dynamic and vulnerable
interdependence. Applied Animal Behaviour Science, 86, 173-204.
Weibull, F. (Oct. 2008). Current issues in imagery used in sports: lectures and seminars.
Jyväskylän Yliopisto.
50
Links
Adidas, miCoach. [22.04.2008, online], http://www.micoach.com
Firstbeat Technologies Ltd. [22.04.2008, online], http://www.firstbeattechnologies.com.
MixMeister Technologies LCC. [15.11.2008, online], BPM Analyzer,
http://www.mixmeister.com/bpmanalyzer/bpmanalyzer.asp.
Nike + iPod. [22.04.2008, online], http://www.apple.com/ipod/nike
Polar Electro Oy. [22.04.2008, online], http://www.polar.fi.
Run to the Beat [10.05.2009, online], http://www.runtothebeat.co.uk
Suunto Oy. [22.04.2008, online], http://www.suunto.com.
Yamaha, Bodibeat. [22.04.2008, online], http://www.yamaha.com/bodibeat/consumer.asp.
(Karageorghis & Priest, 2008)
(Crust & Clough, 2006; Simpson &
Karageorghis, 2006; Karageorghis &
Priest, 2008)
(Karageorghis, Terry & lane, 1999;
Crust, 2004a; Crust, 2004b;
Karageorghis, Priest, Terry,
Chatzisarantis & Lane, 2006; Simpson
& Karageorghis, 2006)
Motivational Quality
(Brunel Music Rating
Inventory)
(Karageorghis & Priest, 2008)
Uplifting Melody &
Harmony
Positive Lyrics
Rhythmic Pattern
(Karageorghis & Priest, 2008;
DeNora, 2000)
Variety in the song
collection
(Karageorghis, Terry & Lane, 1999;
Karageorghis, Priest, Terry,
Chatzisarantis & Lane, 2006;
Karageorghis & Priest, 2008)
Association with Sport,
Exercise, Triumph or
Overcoming Adversity
(North & Hargreaves, 2000)
(Crust & Clough, 2006; Karageorghis
& Priest, 2008)
Musical Preference
(Karageorghis, Terry & lane, 1999;
Crust, 2004b; Karageorghis, Priest,
Terry, Chatzisarantis & Lane, 2006;
Karageorghis & Priest, 2008)
(Iwanaga, 1995; Crust & Clough,
2006; Edworthy & Waring, 2006;
Karageorghis, Jones & Low, 2006)
Rhythmic Energy
Familiarity
(DeNora, 2000; Karageorghis &
Priest, 2008)
Exercise Stage
(Crust & Clough, 2006)
Personality
(Karageorghis & Priest, 2008)
Workout Intensity
(Karageorghis & Priest, 2008)
Type of Activity
Extra-Musical Factors Non-Musical Factors
Tempo
Internal Musical
Factors
51
Attachments
Attachment A: Overview of the preconditions of Optimal Organization with their respective
references.
52
Attachment B: Preceding questionnaire for the participants
Experiment Music and Sport: Preceding Questionnaire
Karolien Dons
1. Personal information
1.1 Name:
1.2 E-mail:
1.3 Year of birth:
1.4 Height:
1.5 Weight:
1.6 Gender:
1.7 Smoker: YES / NO
1.8 Encircle your level of physical activity:
0
1
2
3
4
5
6
0 - Do not participate regularly in programmed
recreation sport or heavy physical activity.
Avoid walking or exertion, e.g. always use
elevator, drive whenever possible instead of
walking.
1 - Do not participate regularly in programmed
recreation sport or heavy physical activity.
Walk for pleasure, routinely use stairs,
occasionally exercise sufficiently to cause
heavy breathing or perspiration.
2 - Participate regularly in recreation or work
requiring modest physical activity, such as
golf, horseback riding, calisthenics,
gymnastics, table tennis, bowling, weight
lifting, yard work for 10 to 60 minutes per
week.
7
7.5
8
8.5
9
4 - Run less than a mile per week or
spend less than 30 minutes per week
in comparative physical activity.
5 - Run 1 to 5 miles per week or spend 30
to 60 minutes per week in
comparative physical activity.
6 - Run 5 to 10 miles per week or spend
to 3 hours per week in comparative
physical activity.
7 - Run over ten miles per week or spend
over than 3 hours per week in
comparative physical activity.
7.5 - Municipal level
8 - Regional / provincial level
3 - Participate regularly in recreation or work
requiring modest physical activity, such as
golf, horseback riding, calisthenics,
gymnastics, table tennis, bowling, weight
lifting, yard work for over an hour per week.
10
8.5 - National level
9 - National championship level
10 - World class level
1.9 Do you have a portable music player (such as mp3-player) that can be used for the
experiment? YES / NO
53
2. Exercise information
2.1 What kind of sports do you do? Indicate also how often you practice.
2.2 How long does your training usually last: … hours … min … sec
2.3 Please describe how your exercise (in case you would run or walk) looks like.
(incl. words like run, warm-up, recovery, cool-down, stretch, jog…).
2.4 Do you sometimes change the training’s:
- Duration yes/no/not applicable
If yes, how?
-
Time of the day yes/no/not applicable
If yes, how?
Intensity yes/no/not applicable
If yes, how?
-
Place yes/no/not applicable
If yes, how?
-
Route yes/no/not applicable
If yes, how?
2.5 Do you use any training devices such as Polar, Suunto, Adidas miCoach, Nike+…?
2.6 Do you listen to music while exercising? YES / NO
2.7 If you (would) use music for exercising, what genre or kind of music (would that be)? You
can encircle several.
Classical
Heavy
Pop
Soul
Other: …
Dance
Hip Hop
Rap
Soundtrack
Experimental
Indie
R&B / Urban
Traditional / Folk
Electronic
Jazz
Rock
World
2.8 Can you name at least 10 examples (max. 15) of the songs you use (or could use) for
exercising?
Please rate by circling also the likeness and the familiarity of the song in the column.
Likeness (1= I don’t like it that much, 6 = I like it very much)
54
Familiarity (1 = I don't know it that much, 6 = I know it very well)
Name of the song + artist
Likeness
Familiarity
1
123456
123456
2
123456
123456
3
123456
123456
5
123456
123456
6
123456
123456
7
123456
123456
8
123456
123456
9
123456
123456
10
123456
123456
11
123456
123456
12
123456
123456
13
123456
123456
14
123456
123456
15
123456
123456
2.9 Please choose 5 songs from your list above and describe what do you feel when you
listen to those songs?
Song nr. … :
Song nr. … :
Song nr. … :
Song nr. … :
Song nr. … :
55
2.10 Would you like to be informed about the outcomes of the research afterwards?
YES / NO
Thank you very much for your time!
Please send or give it back to me (karoliendons@hotmail.com) and please do not hesitate to
ask if something is not clear.
Karolien Dons
karoliendons@hotmail.com
All the information obtained from the experiment will be confident and therefore not
be used for other purposes.
56
Attachment C: Evaluation form after each exercise
Experiment Music and Sport: Evaluation form individual exercise
Karolien Dons
Name:
Number of exercise (1-4):
Date and time of the day:
1.1 Total time of the exercise:
… min … sec
1.2 Please rate (by encircling) the perceived exertion while exercising, namely how heavy
and strenuous the exercise felt for you.
(Note: It depends mostly on the strain and fatigue your muscles feel and on breathlessness
or aches in your chest.)
Use the rating scale hereunder whereby 6 means “no exertion at all” and 20 means “maximal
exertion”.
6
NO EXERTION AT ALL
7
EXTREMELY LIGHT
8
9
10
11
LIGHT
12
13
SOMEWHAT HARD
14
15
HARD (HEAVY)
16
17
VERY HARD
18
19
EXTREMELY HARD
20
MAXIMAL EXERTION
57
1.3 How was your feeling about the performance?
Bad
Could be better
Average
Good
Excellent
1.4 Did anything go wrong during the exercise according to your own standards?
1.5 Did anything go wrong during the exercise according to the experiment’s setup?